These are chat archives for Makuna/NeoPixelBus

14th
Jun 2015
Michael Miller
@Makuna
Jun 14 2015 01:09
@Torxgewinde this inline methods is already inside those, if you glance at the call site (the show method, hard to read) it is wrapped by noInterrupt Interrupt. We have found small changes can cause it be fixed, and then make another change elsewhere and it will brake again.
Michael Miller
@Makuna
Jun 14 2015 06:39
@Torxgewinde seems clI and seI werent updated, these methods call into the library and where not truly disabling all interrupts, nointerrupts does disable ALL interrupts. But maybe these library methods do something else, I will look into deeper.
Michael Miller
@Makuna
Jun 14 2015 17:19
@Torxgewinde and I confirmed, it didn't fix it for my test sample; so you ran into the plaguing issue of small code changes seem to randomly fix it for awhile.
Michael Miller
@Makuna
Jun 14 2015 18:23
Quick poll to ALL: Which Pin do you use to connect your NeoPixels?
sticilface
@sticilface
Jun 14 2015 18:31
Gpio 2 usually on esp 01 and 12.
Michael Miller
@Makuna
Jun 14 2015 21:22
If you have started using the branch NeoPixelAnimator, you can pick up the latest which should have the fix the first pixel problems.
I will try to integrate this into the main branch latter today.
sticilface
@sticilface
Jun 14 2015 21:43
Not yet. I've got much bigger problems. I tried the latest ide, and ota and have made my esp unusable by anything apart from the latest IDE. The ota works great. But when I go back to an older ide it flashes but I get garbage and a load of hex being dumped by serial. Not seen that before. Posted a comment on igrr ota Python script gist page.
Good work though. Did u get to the bottom of it, u were thinking it was a bug?
I'm also having lots of problems with my 'lights sketch' it's got quite unusable at the moment. Crashes a lot. So I need to get that all sorted, before I rewrite the animations method. I will keep u posted though;) and thanks for the hard work!!
Michael Miller
@Makuna
Jun 14 2015 23:48
I believe it is that the instructions has too be pulled from cache, with a flag it can be set to stay in RAM. But while I tried that flag before and it didn't solve it, the issue was that the flag doesn't work in a cpp file, so I reworked the code into a c file and applied the flag and it solves it. remove flag, problem is back.